which severity distinctions have been drawn, such as NRC (1993) and IOM (1997), have tended to combine moderate and severe fluorosis into a single category. The present report focuses more specifically on the severe forms.
The committee compiled prevalence estimates at the person level for severe enamel fluorosis in relation to water fluoride levels from studies around the world. The starting points were the estimates provided in EPA’s documentation supporting the MCLG (50 Fed. Reg. 20164 ) and Appendix C6 of McDonagh et al. (2000a). To these were added results from 24 additional studies (Venkateswarlu et al. 1952; Forsman 1974; Retief et al. 1979; Rozier and Dudney 1981; Subbareddy and Tewari 1985; Haimanot et al. 1987; Kaur et al. 1987; Mann et al. 1987, 1990; Szpunar and Burt 1988; Thaper et al. 1989; Jackson et al. 1995; Cortes et al. 1996; Akpata et al. 1997; Gopalakrishnan et al. 1999; Kumar and Swango 1999; Menon and Indushekar 1999; Rwenyonyi et al. 1999; Sampaio and Arneberg 1999; Awadia et al. 2000; Alarcón-Herrera et al. 2001; Grobler et al. 2001; Ermisş et al. 2003; Wondwossen et al. 2004). Results were excluded if they were for fluorosis indexes other Dean’s index, the TFI, the TSIF, or modifications thereof (e.g., Goward 1982; Nunn et al. 1992); for all fluorosis or for moderate and severe fluorosis combined (e.g., Warnakulasuriya et al. 1992; Mella et al. 1994; Alonge et al. 2000; Burt et al. 2003); for primary or deciduous teeth as opposed to permanent teeth (e.g., McInnes et al. 1982); for different teeth separately with no results at the person level or for all teeth combined (e.g., Opinya et al. 1991); for unbounded upper categories of water fluoride for which no mean or median value was given (e.g., > 1.2 mg/L in Heller et al. , > 2 mg/L in Ray et al. , > 2.5 mg/L in Angelillo et al. ); for bounded but extremely wide water fluoride ranges (e.g., 0.8 to 4.3 mg/L in Haimanot et al. , 0.7 to 4.0 in Beltran-Aguilar et al. , 0.3 to 2.2 mg/L in Wondwossen et al. ). For narrower bounded categories, the midrange water fluoride level was used. Results from studies of children and teenagers (age 20 years or younger) were tallied separately from results for adults. Severe enamel fluorosis was classified as the “severe” classification in Dean’s index and, depending on the groupings created by the original invesgtigators, TFI scores of 4-9 or 5-9 and TSIF scores of 4-7 or 5-7. Because of the wide variability in methods and populations, and the lack of independence when a given study provided more than one result, the estimates were not subjected to formal statistical analyses. Instead, plots of the prevalence estimates in relation to water fluoride concentration were examined for the presence of any clear and obvious patterns or trends.
Figure 4-1 shows 94 prevalence estimates from studies in the United States. Despite the wide range of research methods, fluorosis indexes, water fluoride measurement methods, and population characteristics in these studies conducted over a period spanning half a century, a clear trend is evident.